Research ArticleChemistry

MnTiO3-driven low-temperature oxidative coupling of methane over TiO2-doped Mn2O3-Na2WO4/SiO2 catalyst

See allHide authors and affiliations

Science Advances  09 Jun 2017:
Vol. 3, no. 6, e1603180
DOI: 10.1126/sciadv.1603180


Oxidative coupling of methane (OCM) is a promising method for the direct conversion of methane to ethene and ethane (C2 products). Among the catalysts reported previously, Mn2O3-Na2WO4/SiO2 showed the highest conversion and selectivity, but only at 800° to 900°C, which represents a substantial challenge for commercialization. We report a TiO2-doped Mn2O3-Na2WO4/SiO2 catalyst by using Ti-MWW zeolite as TiO2 dopant as well as SiO2 support, enabling OCM with 26% conversion and 76% C2-C3 selectivity at 720°C because of MnTiO3 formation. MnTiO3 triggers the low-temperature Mn2+↔Mn3+ cycle for O2 activation while working synergistically with Na2WO4 to selectively convert methane to C2-C3. We also prepared a practical Mn2O3-TiO2-Na2WO4/SiO2 catalyst in a ball mill. This catalyst can be transformed in situ into MnTiO3-Na2WO4/SiO2, yielding 22% conversion and 62% selectivity at 650°C. Our results will stimulate attempts to understand more fully the chemistry of MnTiO3-governed low-temperature activity, which might lead to commercial exploitation of a low-temperature OCM process.

  • oxidative coupling of methane
  • ethene
  • heterogeneous catalysis
  • chemical loop
  • oxygen activation
  • methane activation
  • catalyst
  • manganese titanate
  • low-temperature activity
  • natural gas


Methane is a clean and cheap hydrocarbon resource that is abundantly available in natural gas, shale gas, and gas hydrates (1, 2). It can also be sustainably produced from renewable biomass, offering much greater availability than crude oil. In this context, the depletion of global crude oil has stimulated intense efforts on converting methane into high-value chemicals and transportable fuels, which are traditionally derived by petrochemical routes (2, 3). In particular, light olefins, the key building blocks in modern chemical industry, need an urgent shift from crude oil to methane. To date, the industrial-scale conversion from methane to olefins uses an indirect route, where methane is transformed into CO and H2 above 700°C, followed by conversion to methanol and then to olefins (46). Recently, a bifunctional catalyst (ZnCrOx combined with mesoporous SAPO zeolite) (7) and cobalt carbide nanoprisms (8) have been reported to be capable of directly converting CO and H2 into olefins with surprising selectivity under mild conditions. Despite these advances, the strong C–H bonds in methane make this an energy-demanding approach, and such an indirect pathway has a low atom utilization efficiency. One developing technology is the direct conversion of methane into olefins and aromatics in the absence of molecular oxygen (O2) using suitable catalysts. Zeolite-supported Mo catalysts (Mo/zeolites) have been intensively studied (9), and recently, a single-atom iron catalyst embedded in a silica matrix (Fe©SiO2) was reported to have promising methane conversion and light olefin selectivity (10). However, commercial prospects for these processes may be hampered by rapid catalyst deactivation (for Mo/zeolites) and ultrahigh reaction temperature (up to 1100°C for Fe©SiO2).

In principle, methane can also be directly converted into C2 hydrocarbons [ethene (C2H4) and ethane (C2H6)] in the presence of O2 in the so-called oxidative coupling of methane (OCM) (11). The pioneering work by Keller and Bhasin in 1982 initiated a worldwide research effort to explore this process (12). It has been recognized that the OCM process includes both a heterogeneous catalytic step, which involves the activation of O2 and CH4 on the catalyst surface to generate methyl (CH3⋅) radicals (13), and a homogeneous gas-phase step, involving the coupling of CH3⋅ radicals to C2H6 followed by dehydrogenation to C2H4 (3, 11, 14). However, introduction of O2 is double-edged: Although it saves energy to activate CH4, overoxidation is unavoidable. Hundreds of catalysts have been examined since 1982, for C2 selectivity and ability to suppress overoxidation. One early representative catalyst is lithium-doped magnesia (Li/MgO), where [Li+O] centers are formed to effectively generate CH3⋅ from CH4, but rapidly deactivates due to Li loss (15). Another typical class of catalysts is based on lanthanide oxide in both pure and promoted forms (16, 17), whose surface oxygen vacancies are responsible for generating reactive oxygen, but with relatively low C2 selectivity. Of these catalysts, Mn2O3-Na2WO4/SiO2 first reported by Fang et al. (18), is considered optimal in terms of its hundreds of hours of stability and high-temperature productivity (that is, 60 to 80% C2 selectivity and 20 to 30% methane conversion at 800° to 900°C) (19). Since its discovery, this catalyst has been extensively studied with respect to preparation/modification, catalytic mechanism, and microkinetic modeling (1923). Despite these intensive efforts, no catalysts have been successfully applied in a commercial process because of the high reaction temperature (24).

It is widely accepted that the OCM reaction is usually initiated on the solid catalyst surface by reacting CH4 with surface oxygen species to form methyl radicals and continues in the gas phase (19). Activating O2 into desirable species on the catalyst surface is a pivotal step that governs the CH4 activation to CH3⋅ and subsequent oxidative dehydrogenation of C2H6. Recently, Cui et al. provided a highly ordered CaO film doped with Mo2+ ions and suggested that the formed superoxide anions (O2) attribute to CH4 activation (25). Schwach et al. showed that polycrystalline MgO codoped with Fe and Au can effectively activate O2 to peroxy (O22−) species, greatly enhancing the C2 formation with good stability (26). Titanium-containing oxides were also used with the aim to activate O2 into desirable active species, but the C2 yield seemed to be improved only to a limited extent (27, 28). An important question is whether the OCM reaction temperature is dominated by the activation of O2. If so, lowering the O2 activation temperature by catalyst modification will be the key to improving the low-temperature activity of OCM catalysts such as Mn2O3-Na2WO4/SiO2. To test this idea, we explored a TiO2-doped Mn2O3-Na2WO4/SiO2 catalyst for the OCM reaction. MnTiO3 was formed under the reaction conditions and showed a marked enhancement in activity and selectivity at low reaction temperature (that is, catalyst bed temperature): 26% CH4 conversion with 76% C2 plus C3 [C2-C3; propene (C3H6) and propane (C3H8)] selectivity at 720°C or 22% conversion with 62% C2-C3 selectivity at 650°C, using a typical feed of 50% CH4 in air at a total gas hourly space velocity (GHSV) of 8000 ml gcat.−1 hour−1. Moreover, this catalyst is stable for at least 500 hours on stream for the OCM process. Most notably, the in situ–formed MnTiO3 triggers the low-temperature Mn2+↔Mn3+ chemical cycle of O2 activation, thereby leading to improvement in the OCM activity at low temperatures.


Low-temperature active and selective TiO2-doped Mn2O3-Na2WO4/SiO2 catalyst

We initially prepared a series of catalysts by supporting 2.7 weight % (wt %) Mn2O3 and 5.0 wt % Na2WO4 on various Ti-containing materials [including Ti-MWW and TS-1 zeolites with Si:Ti molar ratio of 40:1, pure anatase-TiO2 (a-TiO2), Ti2O3, and perovskite CaTiO3, which all delivered very poor catalytic performance for the OCM reaction (fig. S1)], and the as-prepared catalysts were referred to as Mn2O3-Na2WO4/Ti-MWW (TS-1, a-TiO2, Ti2O3, and CaTiO3). For comparison, a reference catalyst Mn2O3-Na2WO4/SiO2 (2.7 wt % Mn2O3 and 5.0 wt % Na2WO4) was prepared strictly according to the widely applied and recommended procedures (18, 19). The detailed preparation information is given in the Supplementary Materials. All the catalysts were examined in the OCM reaction initially from 800°C (commonly used for the Mn2O3-Na2WO4/SiO2 catalyst), cooling down to 760°C using a feed of 50% CH4 in air at a total GHSV of 8000 ml gcat.−1 hour−1 to quickly screen for a catalyst that might afford acceptable low-temperature catalytic activity and selectivity. The reference catalyst Mn2O3-Na2WO4/SiO2 delivered a performance (Fig. 1A) similar to performances reported at 800°C (1823, 29, 30), indicating its effectiveness in screening these catalysts. On the basis of this reference catalyst, we can separate experimental trials into two groups: Ti-MWW– and TS-1–supported catalysts with higher performance, and pure TiOx- and CaTiO3-supported counterparts with lower performance (fig. S1). Consequently, the Ti-MWW–, TS-1–, and SiO2-supported catalysts were further examined with temperature further decreased from 760° to 720°C; the results are shown in Fig. 1A. The Mn2O3-Na2WO4/Ti-MWW showed the best OCM performance, delivering 26% CH4 conversion and 76% C2-C3 selectivity at 720°C; another zeolite-supported catalyst Mn2O3-Na2WO4/TS-1 delivered slightly lower activity than the Mn2O3-Na2WO4/Ti-MWW catalyst. In contrast, Mn2O3-Na2WO4/SiO2, which is generally considered to be the state-of-the-art catalyst, yielded much lower CH4 conversion and C2-C3 selectivity of only 7.2 and 9.2%, respectively, at 720°C. Thus, the unprecedented low-temperature OCM activity observed on the Ti-MWW– and TS-1–produced catalysts is attributed to the Ti-doping effect.

Fig. 1 The MnTiO3-governed OCM performance of the TiO2-doped Mn2O3-Na2WO4/SiO2 catalysts.

(A) CH4 conversion and C2-C3 selectivity over the Ti-MWW–, TS-1–, and SiO2-supported Mn2O3-Na2WO4 catalysts. (B) XRD patterns of the Mn2O3-Na2WO4/Ti-MWW, Mn2O3-Na2WO4/TS-1, and Mn2O3-Na2WO4/SiO2 catalysts, with (C) the magnified part of 2θ from 31° to 37°. a.u., arbitrary units; α-Crist., α-cristobalite. (D) CH4 conversion and C2-C3 selectivity and (E) XRD patterns as well as (F) the magnified part of 2θ from 31° to 37° of the Mn2O3-Na2WO4/Ti-MWW catalysts with different Si:Ti molar ratio in Ti-MWW zeolites after directly running at 720°C. (G and H) Raman spectra of the Ti-MWW–, TS-1–, and SiO2-supported Mn2O3-Na2WO4 catalysts. Reaction conditions: GHSV of 8000 ml gcat.−1 hour−1 of a feed of 50% CH4 in air. C3 selectivity was 3 to 5%, 2 to 3%, and 0 to 2% for all catalysts at a C2-C3 total selectivity above 60%, between 40 and 60%, and below 40%, respectively.

In situ MnTiO3 formation–dependent low-temperature activity improvement

To optimize the observed Ti-doping effect and exploit it more rationally in future work, it is imperative to explore the origin of the phenomenon, which we accomplished by structural and chemical characterizations of the Mn2O3-Na2WO4/Ti-MWW (Si:Ti of 40:1), Mn2O3-Na2WO4/TS-1 (Si:Ti of 40:1), and Mn2O3-Na2WO4/SiO2 catalysts. Notably, the abovementioned three catalysts (that is, first reacting at 800°C for 2 hours and then at 720°C for 0.5 hour; see detailed treatment history in table S1) were exclusively denoted as Cat-Ti-MWW for Mn2O3-Na2WO4/Ti-MWW, Cat-TS-1 for Mn2O3-Na2WO4/TS-1, and Cat-SiO2 for Mn2O3-Na2WO4/SiO2. First, the elemental analyses and low-temperature N2-sorption measurements of these three catalysts revealed almost identical content of Mn, W, and Na as well as similar specific surface areas (1.2 to 1.4 m2 g−1) (table S2 and fig. S2), indicating that the low-temperature activity difference should be correlated with their different phase or chemical properties rather than element content and surface area. Subsequently, the phase properties of the Cat-Ti-MWW, Cat-TS-1, and Cat-SiO2 were probed by x-ray diffraction (XRD). The XRD patterns are shown in Fig. 1B, and the magnified 2θ part from 31° to 37° is shown in Fig. 1C. For the reference catalyst Cat-SiO2, the phase of the main α-cristobalite as well as of Na2WO4 and Mn2O3, which is commonly reported, was detected (1823); the Cat-Ti-MWW and Cat-TS-1 also had α-cristobalite and Na2WO4 phases. There was a phase transformation from Mn2O3 to a new compound, MnTiO3, with almost full transformation for the Cat-Ti-MWW and partial for the Cat-TS-1 (Fig. 1C). This observation, combined with the higher activity/selectivity over the Cat-Ti-MWW and Cat-TS-1, indicates an improvement effect similar to that of MnTiO3 on the CH4 low-temperature conversion. As noted above, after undergoing OCM reaction at 800°C before performing at 720°C, MnTiO3 was in situ–formed over these three catalysts. Conversely, MnTiO3 was formed only for the Ti-MWW–supported catalyst but not for the TS-1–supported catalyst when directly running at 720°C, associated with a CH4 conversion of 25% and a C2-C3 selectivity of 73% for the Ti-MWW–supported catalyst but a conversion of only 4% and a C2-C3 selectivity of 16% for the TS-1–supported catalyst (fig. S3). Once MnTiO3 was formed for the TS-1–supported catalyst after reaction at 800°C, a high CH4 conversion of 23% could be obtained with C2-C3 selectivity increased to 69% as the reaction temperature was decreased to 720°C. The abovementioned results demonstrate that the appearance of MnTiO3 is responsible for the low-temperature CH4 conversion (fig. S3).

To confirm the as-observed MnTiO3-governed low-temperature activity improvement, we synthesized a series of Ti-MWW zeolites with varying Ti:Si ratio from 1:80 to 1:40, as well as a full-Si zeolite as a reference (fig. S4A); all the Ti-MWW–supported Mn2O3-Na2WO4 catalysts were directly tested at 720°C. With increase of the Ti:Si ratio in the zeolites, the catalysts showed a monotonously increasing CH4 conversion and a synchronous MnTiO3 intensity (Fig. 1, D to F). Moreover, TS-1 zeolites with varying Ti:Si ratios were also synthesized (fig. S4B), and their supported catalysts were also tested in the OCM reaction. Running all these TS-1–supported catalysts directly at 720°C did not lead to MnTiO3 and was associated with a very low CH4 conversion of <5% and a C2-C3 selectivity of <20%. However, MnTiO3 was formed after the reaction at 800°C and could be sustained when reaction temperature was decreased to 720°C, and CH4 conversion was also progressively increased at 720°C along with the corresponding amount of MnTiO3 (fig. S4, C to E). These results provide solid proof for the critical role played by MnTiO3 in enhancing catalyst performance, although the slight differences, which likely result from the structural rigidity difference between Ti-MWW and TS-1 zeolites, were not explored further.

Despite the clear proof from XRD characterization, the question naturally arises as to the surface composition and structure of the Cat-Ti-MWW and Cat-TS-1 compared with the Cat-SiO2 because the catalytic reaction is a surface-dependent process. To answer this question, we further probed the surface states of these three catalysts by the surface-sensitive techniques of x-ray photoelectron spectroscopy (XPS) and Raman spectroscopy as good complements for XRD. The XPS results show almost identical surface content of W, Mn, and Na for these catalysts (table S3), indicating that their activity/selectivity should be controlled by other factors rather than the surface element contents. However, it should be noted that the XPS spectrum of Mn in Mn2O3 is identical to that in MnTiO3 (fig. S5A); therefore, the Mn2O3-to-MnTiO3 evolution cannot be distinguished by XPS. Moreover, the binding energies of W, Mn, and Na species on the Cat-Ti-MWW and Cat-TS-1 shifted very slightly compared with the Cat-SiO2 surface (fig. S5, B to F), indicating that the MnTiO3 formation did not markedly modify the original electronic states of these species on the catalyst surface.

Another surface-sensitive technique of Raman spectroscopy was used because it is sensitive to the local structure of oxides especially with poor crystallinity (31, 32). For the Cat-Ti-MWW, Cat-TS-1, and Cat-SiO2, Mn2O3 and Na2WO4, which are generally considered to be crucial for the traditional SiO2-supported catalyst in the OCM reaction (19), and MnTiO3, which is paramount for low-temperature activity improvement as evidenced by the above experiments, were analyzed in particular. For the Cat-SiO2 catalyst, we observed signals of Mn2O3 and Na2WO4 as well as α-cristobalite (Fig. 1G and fig. S6) that were identical to those reported previously (33). On the Cat-Ti-MWW and Cat-TS-1, in contrast, we detected no signals of Mn2O3, but found strong signals of MnTiO3, and clear signals of Na2WO4 and α-cristobalite (Fig. 1G and fig. S6). These Raman results also showed a clear transformation from Mn2O3 to MnTiO3 on the Cat-Ti-MWW and Cat-TS-1, which confirmed again the critical role of MnTiO3 in improving low-temperature activity. Quite different temperature-dependent surface evolution was also captured by Raman for the Ti-MWW– and TS-1–supported catalysts (Fig. 1H). When directly running at 720°C for 0.5 hour, Mn species in the Ti-MWW–supported catalyst were almost fully transformed into MnTiO3 associated with a CH4 conversion of 25%, whereas the TS-1–supported catalyst displayed a spectrum identical to that of TiO2 with a conversion of 4%. As expected, after the TS-1–supported catalyst underwent the OCM reaction at 800°C for 2 hours in advance, dominant MnTiO3 was detected, associated with a sharp increase of CH4 to 23% at 720°C (fig. S3). The Raman-indicated correspondence of the high CH4 conversion to the appearance of surface MnTiO3 corroborated the similar improvement effect of MnTiO3 on the low-temperature CH4 conversion once again.

Synergistic catalysis of MnTiO3 with Na2WO4

Our probes of phase and surface structures consistently showed the critical role of MnTiO3 in enhancing the low-temperature OCM performance of the TiO2-doped Mn2O3-Na2WO4/SiO2 catalyst; even more notably, CH4 conversion and C2-C3 selectivity were progressively enhanced with increasing amount of MnTiO3 (Fig. 1, D to F, and fig. S4), suggesting that MnTiO3 contains the active sites for the OCM reaction. To verify this point, we prepared and examined an MnTiO3/SiO2 catalyst in the OCM reaction. At 800°C, this catalyst offered slightly lower CH4 conversion (22% versus 26%) and much lower C2-C3 selectivity (47% versus 75%) than the Mn2O3-Na2WO4/SiO2 catalyst, but similar at 760°C (17% versus 14%; 34% versus 38%) and much higher at 720°C (15% versus 7.2%; 27% versus 9.2%) (fig. S7). Nevertheless, both the activity and the selectivity of the MnTiO3/SiO2 catalyst was much lower than the results of the Mn2O3-Na2WO4/Ti-MWW with MnTiO3-Na2WO4 combination (26 to 28% conversion and 76 to 79% C2-C3 selectivity at 720° to 800°C; Fig. 1A). Moreover, two other catalysts, Mn2O3/SiO2 and Na2WO4/SiO2, were prepared and tested, delivering much lower CH4 conversion (8 to 14% for Mn2O3/SiO2 and 5 to 10% for Na2WO4/SiO2) and C2-C3 selectivity (14 to 18% for Mn2O3/SiO2 and 28 to 33% for Na2WO4/SiO2) at 720° to 800°C (fig. S7). It is thus rational to infer that the coexistence of Na2WO4 and MnTiO3 is paramount for the low-temperature OCM reactivity improvement. Subsequently, the question arises whether the MnTiO3-Na2WO4 combination affects their physical structure [such as solely improving dispersion to enhance CH4 conversion (34)] or the intrinsic catalytic performance. Therefore, a kinetic study was carried out over the Mn2O3-Na2WO4/SiO2 and Mn2O3-Na2WO4/Ti-MWW catalysts, and the apparent activation energies (Ea) were calculated with the results as shown in fig. S8. Mn2O3-Na2WO4/Ti-MWW provided a much lower Ea (80 to 110 kJ/mol) than Mn2O3-Na2WO4/SiO2 (180 to 210 kJ/mol) (see the Supplementary Materials for the calculation details), indicating that there are different active sites in the two catalysts and that the MnTiO3-Na2WO4 combination markedly promotes the catalyst intrinsic performance.

MnTiO3-triggered Mn2+↔Mn3+ chemical cycle—The nature of low-temperature OCM catalysis

As mentioned above, it has been recognized that the OCM process obeys the mechanism of heterogeneous-homogeneous radical reactions (3, 11, 14), and the heterogeneous surface activation of O2 and CH4 is the pivotal prerequisite for the OCM process. Therefore, we inferred that the Mn2O3-Na2WO4/Ti-MWW catalyst can activate O2 and CH4 more effectively than the other catalysts. To confirm this conjecture, we designed and conducted a three-step experiment over the Mn2O3-Na2WO4/Ti-MWW and Mn2O3-Na2WO4/SiO2 catalysts: reducing in CH4 stream for 0.5 hour and then switching to an O2 stream, followed by switching to CH4 at the desired temperature for some time (see detailed treatment history in table S1). The evolution of phase structures and surface states of these as-resulting catalysts was characterized by XRD and Raman techniques, with results as shown in Fig. 2. After reducing in CH4 at 800°C, Mn species for the Mn2O3-Na2WO4/Ti-MWW catalyst existed in the form of the MnTiO3 phase; then, full oxidation of MnTiO3 to form Mn2O3 took only 1 min in the O2 stream, and the subsequent reversal of Mn2O3 into MnTiO3 took 10 min in the CH4 stream at 800°C (Fig. 2A). Differently, the Mn2O3-Na2WO4/SiO2 catalyst after reducing in the CH4 stream at 800°C displayed clear MnWO4 diffractions. Thereafter, full oxidation of MnWO4 to form Mn2O3 took 3 min in the O2 stream, whereas the reversal of Mn2O3 into MnWO4 took 15 min in the CH4 stream at 800°C (Fig. 2A). At 760°C, the evolution of the MnTiO3↔Mn2O3 cycle slowed down slightly for the Mn2O3-Na2WO4/Ti-MWW catalyst, but that of the MnWO4↔Mn2O3 cycle slowed down greatly for the Mn2O3-Na2WO4/SiO2 catalyst (Fig. 2B). The oxidation of MnTiO3 to form Mn2O3 took 2 min, and the reversal of Mn2O3 to MnTiO3 took 15 min. However, the transition of MnWO4 to Mn2O3 took a much longer time of 15 min, and that from Mn2O3 into MnWO3 also took a much longer time of 45 min. At 720°C, the MnTiO3↔Mn2O3 cycle could easily take place despite a longer time: MnTiO3 to Mn2O3 took 4 min and Mn2O3 to MnTiO3 took 30 min (Fig. 2C). Most notably, MnWO4 to Mn2O3 took 30 min, but Mn2O3 could not be fully reversed into MnWO4 even after 1.5 hours in a CH4 stream at 720°C (Fig. 2C). Obviously, the Mn2O3-Na2WO4/Ti-MWW catalyst delivered a much easier Mn2+↔Mn3+ cycle especially at 720°C (Fig. 2, D to F) because of the participation of MnTiO3. Combining the OCM results of these two catalysts, we conclude that the more facile Mn2+↔Mn3+ cycle of the Mn2O3-Na2WO4/Ti-MWW catalyst matches neatly with its much higher performance at 720°C. In addition, the Raman results of the Mn2O3-Na2WO4/Ti-MWW and Mn2O3-Na2WO4/SiO2 catalysts (Fig. 2, G to I) corroborated that the Mn2O3-Na2WO4/Ti-MWW catalyst offered a much easier Mn2+↔Mn3+ cycle, implying a much higher activity to activate O2 (via one half-cycle of Mn2+→Mn3+) and to seize the reducing species (for the other half-cycle of Mn3+→Mn2+) than the Mn2O3-Na2WO4/SiO2 catalyst at low temperature.

Fig. 2 The temperature-dependent evolution of phase structures and surface states for the Ti-MWW– and SiO2-supported Mn2O3-Na2WO4 catalysts in O2 or CH4 at 720° to 800°C.

(A to C) XRD patterns. The Mn3+ fractions evolving at (D) 800°C, (E) 760°C, and (F) 720°C. (G to I) Raman spectra. The Ti-MWW–supported catalyst reduced in CH4 at 800°C for 0.5 hour exhibits dominant MnTiO3 signal. After subsequent oxidation in O2 stream for 1 min at 800°C, 2 min at 760°C, and 4 min at 720°C, the MnTiO3 signal disappears, whereas the TiO2 appears, and is reformed in CH4 stream for 9 min at 800°C, 15 min at 760°C, and 30 min at 720°C. The SiO2-supported catalyst reduced in CH4 at 800°C for 0.5 hour exhibits a dominant MnWO4 signal. After subsequent oxidation in O2 stream for 3 min at 800°C, 15 min at 760°C, and 30 min at 720°C, the MnWO4 signal disappears, whereas the Mn2O3 appears, and can be reformed in CH4 stream for 15 min at 800°C, 35 min at 760°C, and 60 min at 720°C. The detailed calculations of the Mn3+ fractions are given in the Supplementary Materials.

Conflicting reports on the structure of the active sites of the Mn2O3-Na2WO4/SiO2 catalyst are found in literature, and three typical descriptions have been widely proposed (19). The unitary site of Na−O−Mn was suggested because of the similarity between the catalytic performance of Mn2O3-Na2WO4/MgO and Na2MnO4/MgO. In contrast, a redox mechanism involving a W6+/W4+ couple with W−O−Si bonds was suggested, and the gas-phase O2 was proposed to be involved in electron transfer. In particular, a two-metal site model was proposed, where O2 is activated on an Mn3+ site to take charge of catalyst activity and CH4 is activated on the neighboring W6+ site to increase catalyst selectivity, with oxygen spillover from Mn2O3 to the Na2WO4 to accomplish the reactive cycle. Our results indicate that the Mn2+↔Mn3+ cycle is confined to Mn2O3 and MnWO4 (Fig. 2, A to C); therefore, we prefer the Mn−W combined site model for the Mn2O3-Na2WO4/SiO2 catalyst in our present OCM reaction. Moreover, the evolution rate of the Mn2+↔Mn3+ cycle in Mn2O3-MnWO4 was much lower than in MnTiO3-Na2WO4, especially at low temperatures (Fig. 2); therefore, we hypothesized that MnTiO3 plays a dominant role in enhancing low-temperature activity and Na2WO4 is key to improving C2-C3 selectivity. To further confirm this conclusion, we prepared four catalysts, including MnWO4-Mn2O3, MnTiO3-Mn2O3, MnWO4-Mn2O3-Na2WO4, and MnTiO3-Mn2O3-Na2WO4, by grinding. As expected, the catalysts containing MnTiO3 delivered a much higher CH4 conversion at 720° to 760°C than the counterparts with absence of MnTiO3, and the catalysts with Na2WO4 delivered a much higher C2-C3 selectivity at 720° to 800°C (fig. S9). For example, the MnTiO3-Mn2O3 catalyst delivered a CH4 conversion of 16%, whereas Mn2O3-MnWO4 delivered only 4% at 720°C, but selectivities were very low for both catalysts (42% for MnWO4-Mn2O3 and 29% for MnTiO3-Mn2O3). As we added Na2WO4 into MnTiO3-Mn2O3, C2-C3 selectivity was sharply increased from 29 to 73%, whereas CH4 conversion was slightly increased from 16 to 22%. These experimental results clearly demonstrate that MnTiO3 was mainly responsible for low-temperature O2 and CH4 activation and Na2WO4 for increased selectivity, indicating that their combination (Na2WO4-MnTiO3) resulted in a much higher conversion rate and selectivity. This observation could be tentatively attributed to the synergetic interactions between Mn species (Mn2O3 or MnTiO3) and Na2WO4 [see temperature-programmed reduction (TPR) results in fig. S10 and the corresponding discussion in the Supplementary Materials], in which the detailed spectroscopic/microscopic investigations are particularly desirable.

On the basis of the evolution behavior from Mn2+ to Mn3+ (associated with MnTiO3 or MnWO4 to Mn2O3 for O2 activation) of the Mn2O3-Na2WO4/Ti-MWW and Mn2O3-Na2WO4/SiO2 catalysts in the O2 stream at 720° to 800°C (Fig. 2, D to F), the transition rate from Mn2+ to Mn3+ was calculated and shown in Fig. 3A (see the Supplementary Materials for the calculation details). At 800°C, the transition rate from Mn2+ to Mn3+ was 98%/min for the Mn2O3-Na2WO4/Ti-MWW catalyst and 53%/min for the Mn2O3-Na2WO4/SiO2 catalyst, associated with similar CH4 conversion (26 to 28%) and C2-C3 selectivity (76 to 79%). At 760°C, the rate from Mn2+ to Mn3+ for the Mn2O3-Na2WO4/Ti-MWW was 37%/min with a CH4 conversion of 27% and a C2-C3 selectivity of 78%, whereas the rate for Mn2O3-Na2WO4/SiO2 was decreased to only 7%/min with a low CH4 conversion of 14% and a poor C2-C3 selectivity of 38%. At 720°C, most notably, the rate from Mn2+ to Mn3+ for the Mn2O3-Na2WO4/Ti-MWW was still retained at 13%/min with a high CH4 conversion of 26% and a high C2-C3 selectivity of 76%, but a very slow rate of only 4%/min was obtained for the Mn2O3-Na2WO4/SiO2 with only 7.2% CH4 conversion and 9.2% C2-C3 selectivity. These results show that [Mn2O3-Na2WO4]–based catalyst performance was strongly dependent on the Mn2+-to-Mn3+ rate independent of temperature, and a transition rate of 10%/min seemed critical to increased OCM conversion/selectivity. As previously shown for the Mn2O3-Na2WO4/SiO2 catalyst, O2 is activated on the Mn2+ site and CH4 on the W6+ site, and the active oxygen species play a key role in activating CH4 after spillover to the W6+ site (19). Accordingly, it is safe to say that the Mn2O3-Na2WO4/SiO2 catalyst has an adequate O2 activation rate only through the Mn2O3↔MnWO4 chemical cycle at 800°C and therefore delivers acceptably high OCM performance only at 800°C (Fig. 3, A and B). For the Mn2O3-Na2WO4/Ti-MWW catalyst, in contrast, because of the MnTiO3-triggered low-temperature Mn2+↔Mn3+ chemical cycle, a high O2 activation rate could be obtained at lower temperature (for example, 720°C); as a result, this catalyst yielded better low-temperature OCM performance (Fig. 3, A and B).

Fig. 3 The Mn2+-to-Mn3+ transition rate and proposed catalytic recycle for OCM process.

(A) Temperature-dependent transition rate of Mn2+ to Mn3+ correlated to CH4 conversion. (B) The Mn2O3-Na2WO4/Ti-MWW and Mn2O3-Na2WO4/SiO2 catalysts and the proposed catalytic cycles of Mn2O3-Na2WO4 and MnTiO3-Na2WO4 combinations. Detailed calculations of the transition rate are given in the Supplementary Materials.

A practical MnTiO3-Na2WO4/SiO2 catalyst

Inspired by the obtained insight into the MnTiO3-enhanced low-temperature catalysis for the OCM process, we successfully prepared a catalyst by a solvent-free method, simply grinding commercial Mn2O3, TiO2, Na2WO4, and SiO2 gel in a high-speed ball mill. By varying the loadings of Mn2O3, TiO2, and Na2WO4, the mixed system yielded a highly effective OCM catalyst in the range of 8 to 15 wt % for Na2WO4, and 6 to 28 wt % for Mn2O3 plus TiO2 (with a stoichiometric ratio of Mn2O3 to TiO2 to be fully transformed into MnTiO3), with the SiO2 gel making up the balance (fig. S11). For example, the 6.2Mn2O3-6.3TiO2-10Na2WO4/SiO2 catalyst (6.2 wt % Mn2O3, 6.3 wt % TiO2, and 10 wt % Na2WO4) was first activated in the reaction stream at 800°C for 2 hours and delivered an interesting CH4 conversion of 24% with 73% C2-C3 selectivity at 720°C (comparable to the conversion/selectivity for the Mn2O3-Na2WO4/Ti-MWW catalyst at 720°C in Fig. 1A) and a more interesting 22% CH4 conversion with 62% C2-C3 selectivity even at 650°C (Fig. 4A). After the reaction from 800° to 650°C, TiO2 and Mn2O3 were in situ–transformed into MnTiO3 (Fig. 4, B and C) with the corresponding loading of 11.8 wt %. The catalytically initiated reaction over this 11.8MnTiO3-10Na2WO4/SiO2 catalyst (that is, derived from 6.2Mn2O3-6.3TiO2-10Na2WO4/SiO2 after undergoing OCM reaction at 800° to 650°C) compared very favorably with other reported OCM catalysts, as summarized in table S4. Among them, an ordered mesoporous SBA-15–supported Mn2O3-Na2WO4 catalyst (34) has been recently reported to yield better results than the reference catalyst Mn2O3-Na2WO4/SiO2 at 750°C: CH4 conversion, 14% versus 7%; C2-C3 selectivity, 63% versus 50%. By comparison, our 11.8MnTiO3-10Na2WO4/SiO2 catalyst delivered a much better low-temperature performance than these reported catalysts for the OCM process (table S4). Besides the enhanced low-temperature activity and selectivity, such 11.8MnTiO3-10Na2WO4/SiO2 catalyst provided promising chemical/mechanical stability during the OCM reaction. With a total GHSV of 8000 ml gcat.−1 hour−1, a scale-up experiment using 10 ml of particulate catalyst (20 to 40 meshes) was carried out for the stability test, and no deactivation was observed during a 500-hour run at 720°C with a feed gas of 50% methane in air (Fig. 4D). CH4 conversion remained at 22 to 25% throughout the entire 500-hour testing, whereas C2-C3 selectivity was retained at 68 to 73% with an ethylene/ethane ratio of ~1.9. Similar behavior was also seen at 800°C, with well-retained CH4 conversion at 24 to 28% and C2-C3 selectivity at 73 to 77% for a 400-hour run (fig. S12). Moreover, our 6.2Mn2O3-6.3TiO2-10Na2WO4/SiO2 catalyst showed a marked low-temperature reaction ignition property that is an important consideration for the practical OCM process. The OCM reaction could be directly started over this catalyst at a low reaction temperature (that is, catalyst bed temperature) of 650°C, offering a high conversion of 17% but a low C2-C3 selectivity of only 47%. As the reaction temperature was increased from 650° to 720°C, a high C2-C3 selectivity of 73% could be obtained with an improved conversion of 25% (fig. S13A) because of the facilitation of in situ formation of MnTiO3 at that high temperature point. The MnTiO3 phase was detected on the sample after direct reaction at 720°C for only 1 hour and remained almost unchanged with prolonged time on stream. In contrast, only a small amount of MnTiO3 was formed after a direct reaction at 650°C for 1 hour; however, its content increased with prolonged time on stream, after 8 hours becoming comparable in results to those of a direct reaction at 720°C for 1 hour (fig. S13B). Not surprisingly, at 650°C, CH4 conversion and C2-C3 selectivity gradually increased from 17 and 47% at the beginning to 21 and 60% after 8 hours on the reaction stream in association with clear MnTiO3 phase formation while remaining almost stable along with further prolonged time on stream to 12 hours (fig. S13B). Moreover, because of the similarity of MnTiO3 phase intensity between samples after direct reaction at 720°C for 1 hour and those after direct reaction at 650°C for 8 hours or longer (fig. S13B), the former catalyst sample, not surprisingly, delivered results comparable to those of the latter one, as the reaction temperature was reduced from 720° to 650°C (fig. S13).

Fig. 4 The OCM performance and XRD/Raman results of the 6.2Mn2O3-6.3TiO2-10Na2WO4/SiO2 catalyst.

(A) Temperature-dependent CH4 conversion and C2-C3 selectivity. (B) XRD pattern and (C) Raman spectrum of this catalyst. (D) CH4 conversion and C2-C3 selectivity along with the time on stream at 720°C. Reaction conditions: GHSV of 8000 ml gcat.−1 hour−1 of a feed of 50% CH4 in air. The C3 selectivity was 3 to 5%, 2 to 3%, and 0 to 2% for all catalysts at a C2-C3 total selectivity above 60%, between 40 and 60%, and below 40%, respectively.


Our results established a TiO2-doped Mn2O3-Na2WO4/SiO2 catalyst system, which provides an enhanced low-temperature activity and selectivity in combination with promising stability for the OCM process. MnTiO3 is in situ–generated in the reaction stream and triggers the low-temperature MnTiO3↔Mn2O3 chemical cycle for O2 activation, thereby leading to a marked improvement of the low-temperature activity/selectivity. Guided by these findings, we discovered an MnTiO3-Na2WO4/SiO2 catalyst obtained by the ball milling method using Mn2O3, TiO2, Na2WO4, and SiO2 gel as starting materials, of which the reaction temperature can be further lowered to 650°C with acceptable CH4 conversion (~22%) and C2-C3 selectivity (~62%). We suggest that the reaction temperature might be decreased further if the Mn2+↔Mn3+ chemical cycle could be accelerated at much lower temperatures for the [Mn2O3-Na2WO4]–based catalysts. To accomplish this goal, discovery of new Mn2+↔Mn3+ chemical cycles and a deep understanding of the chemistry of the MnTiO3 (or analog)–governed low-temperature OCM activity/selectivity at the atomic level are particularly desirable. In addition, yield or selectivity of C2-C3 products will need to be further improved, perhaps in an appropriate reactor concept such as a membrane and/or structured reactor.


Catalyst preparation

The Mn2O3-Na2WO4/support catalysts with Mn2O3 loading of 2.7 wt % [Mn(NO3)2 aqueous solution as precursor] and Na2WO4 loading of 5 wt % (Na2WO4·2H2O as precursor) for the OCM reaction were obtained by the incipient wetness impregnation (IWI) method (18, 19). The supports, including Ti-MWW and TS-1 zeolites (Si:Ti molar ratio of 40:1 to 80:1), amorphous SiO2 gel (~73 μm), pure a-TiO2 (~44 μm, 99.8%), Ti2O3 (~44 μm, 99.8%), and CaTiO3 (~44 μm, 99%+), were impregnated with an aqueous solution of Mn(NO3)2 [analytical reagent grade (AR), Sinopharm Chemical Reagent Co.] and Na2WO4·2H2O (AR, Sinopharm Chemical Reagent Co.) containing appropriate concentrations at room temperature, followed by constant stirring for 5 hours at room temperature and for 1 hour at 180°C. The resulting slurry was dried at 100°C overnight and then calcined in air at 550°C for 2 hours. The Ti-MWW and TS-1 zeolites (Si:Ti molar ratio of 40:1 to 80:1) were synthesized according to the reported methods (35, 36). The amorphous SiO2 gel and a-TiO2 were purchased from Aladdin Industrial Corporation. The Ti2O3 and CaTiO3 were purchased from Alfa Aesar Co. Ltd.

The MnTiO3/SiO2, Mn2O3/SiO2, and Na2WO4/SiO2 catalysts were also prepared by the IWI method. The Mn, Na, and W contents of these three catalysts were consistent with the abovementioned loadings. Taking the MnTiO3/SiO2 catalyst as an example, the amorphous SiO2 gel was mixed with a-TiO2 (Si:Ti molar ratio of 40:1) and impregnated with an aqueous solution of Mn(NO3)2 (AR, Sinopharm Chemical Reagent Co.). The solution system was stirred for 5 hours at room temperature and for 1 hour at 180°C. The obtained slurry was dried at 100°C overnight and then calcined in air at 550°C for 2 hours.

The MnWO4-Mn2O3 (5:4 weight ratio), MnTiO3-Mn2O3 (1:1 weight ratio), MnWO4-Mn2O3-Na2WO4 (5:4:1 weight ratio), and MnTiO3-Mn2O3-Na2WO4 (5:5:1 weight ratio) catalysts were prepared by grinding in a high-energy planetary ball mill (37). These four catalysts were milled for 2 hours to obtain a homogeneous mixture. The mass ratio of the balls to particles was 10:1, and the rotation speed was 320 rpm.

The practical catalyst Mn2O3-TiO2-Na2WO4/SiO2 was also prepared by grinding (38) in a high-energy planetary ball mill for 2 hours with the same mass ratio of balls to particles and rotation speed as for the other catalysts. By varying the contents of Mn2O3, TiO2, and Na2WO4, the mixed system led to a highly effective OCM catalyst in the range of 8 to 15 wt % loading for Na2WO4 and 6 to 28 wt % loading for Mn2O3 plus TiO2 (with proper Mn/Ti ratio to form MnTiO3), with amorphous SiO2 gel making up the balance.


The catalysts were characterized by scanning electron microscopy (SEM; Hitachi S-4800) equipped with an energy-dispersive x-ray fluorescence spectrometer (EDX; Oxford), inductively coupled plasma atomic emission spectrometry (ICP Thermo IRIS Intrepid II XSP), and XPS [ESCALAB 250Xi spectrometer, Al Kα, adventitious C 1s line (284.8 eV) as the reference]. The specific surface area was determined using standard Brunauer-Emmett-Teller theory on the N2 adsorption isotherm measured at −196°C on a Quantachrome Autosorb-3B instrument. The pore size distribution was determined using the Barrett-Joyner-Halenda (BJH) method. TPR with hydrogen (H2-TPR) was performed on a Quantachrome ChemBET 3000 chemisorption apparatus with a thermal conductivity detector (TCD). In each H2-TPR experiment, the sample (100 mg) purged by He at 300°C for 1 hour in advance was heated from 20° to 1000°C in a gas mixture of 5% H2 in N2 (30 ml min−1) at a rate of 10°C min−1. XRD was performed on a Rigaku Ultima IV diffractometer with Cu Kα radiation (35 kV and 25 mA). The Raman measurements were carried out using a Raman spectrometer (Renishaw inVia) with a 532-nm semiconductor laser as excitation, equipped with a charge-coupled device camera enabling microanalysis on a sample point. The scanning range was set from 80 to 2000 cm−1.

Reactivity tests

The OCM reaction was performed in a fixed bed quartz tube reactor, 400 mm of straight cylindrical tubing with an internal diameter of 16 mm, under atmospheric pressure. The catalyst bed was placed between quartz wool plugs in the reactor. For the Ti-MWW–supported catalyst, 0.25 g of catalyst was loaded in the reactor with the catalyst bed thickness of approximately 10 mm; however, the density of every other catalyst is much higher than that of the Ti-MWW–supported catalyst, and to get the identical catalyst bed thickness of about 10 mm, every other catalyst of 1.5 g was loaded in the reactor. The reactants, CH4 (99.99%) and O2 (99.999%) with dilution of N2 (99.99%), were cofed into the reactor using calibrated mass flow controllers. The CH4:O2:N2 molar ratio of 5:1:4 imitated the contents of a coal bed gas (50 volume % CH4 in air) with a GHSV of 8000 ml gcat.−1 hour−1. The reaction temperature (that is, catalyst bed temperature) was monitored by a thermocouple placed in the middle of the catalyst bed and was 720°, 740°, 760°, 780°, and 800°C. The effluent gas was analyzed with an online gas chromatograph equipped with a TCD using a 60-m DM-Plot Msieve 5A column (for the separation of N2, O2, CO, and CH4) and a 30-m DM-Plot Q capillary column (for the separation of CO2, CH4, C2H4, C2H6, C3H6, and C3H8) in parallel. The CH4 conversion (CCH4) and C2H4/C2H6/C3H6/C3H8 selectivity (SC2–3) were calculated using the standard normalization method on the basis of carbon atom balance and defined as follows (Eqs. 1 and 2)Embedded Image(1)Embedded Image(2)

No formation of carbon deposit was observed. The desired products of the OCM reaction were C2H4, C2H6, C3H6, and C3H8. Their selectivity was described as the C2-C3 selectivity. Reaction data at each reaction condition were collected after running for at least 0.5 hour.


Supplementary material for this article is available at

Supplementary Text

fig. S1. CH4 conversion and C2-C3 selectivity for the pure supports and the supported Mn2O3-Na2WO4 catalysts.

fig. S2. SEM and EDX mapping images.

fig. S3. XRD patterns and testing results of the 2.7Mn2O3-5.0Na2WO4/Ti-MWW and 2.7Mn2O3-5.0Na2WO4/TS-1 catalysts under different reaction conditions.

fig. S4. XRD patterns and testing results for various samples with different Si:Ti molar ratio (or Ti content).

fig. S5. XPS spectra of various samples.

fig. S6. Raman spectra of various samples.

fig. S7. Testing results of the catalysts with different active components.

fig. S8. Ea calculations.

fig. S9. Testing results of the catalysts with different active components prepared by the grinding method.

fig. S10. H2-TPR profiles and XRD patterns.

fig. S11. Effects of Mn2O3 plus TiO2, and Na2WO4 loadings on the OCM performance for the Mn2O3-TiO2-Na2WO4/SiO2 catalyst.

fig. S12. Stability testing of the 6.2Mn2O3-6.3TiO2-10Na2WO4/SiO2 catalyst.

fig. S13. Testing results and XRD patterns for the 6.2Mn2O3-6.3TiO2-10Na2WO4/SiO2 catalyst under different reaction conditions.

table S1. Detailed treatment history of some catalysts for XRD and Raman measurements.

table S2. Specific surface areas and real contents of Mn, W, Na, and Ti of all used catalysts.

table S3. Surface contents of W, Mn, Na, Ti, Si, O, and C measured by XPS for the used catalysts.

table S4. CH4 conversion and C2-C3 selectivity over representative catalysts.

References (3849)

This is an open-access article distributed under the terms of the Creative Commons Attribution-NonCommercial license, which permits use, distribution, and reproduction in any medium, so long as the resultant use is not for commercial advantage and provided the original work is properly cited.


Acknowledgments: We acknowledge P. Wu and Y. Liu at the East China Normal University for their assistance in supplying Ti-MWW and TS-1 zeolite samples. We also thank R. Prins from the ETH-Zürich for the helpful discussion. Funding: This work has been funded by the National Natural Science Foundation of China (21473057, U1462129, 21322307, 21273075, and 21076083) and the “973 Program” (2011CB201403) from the Ministry of Science and Technology of the People’s Republic of China. Author contributions: Y.L. and G.Z. conceived the idea for the project. P.W. and Y.W. conducted the material synthesis. P.W. performedthe structural characterizations and catalytic test. Y.L., G.Z., and P.W. discussed the catalytic results. Y.L. and G.Z. drafted the manuscript. P.W. and G.Z. contributed equally to this work. All authors discussed and commented on the manuscript. Competing interests: Y.L., P.W., and G.Z. have a patent application related to this work filed with the Chinese Patent Office on 13 March 2017 (201710146190.1). All other authors declare that they have no competing interests. Data and materials availability: All data needed to evaluate the conclusions in the paper are present in the paper and/or the Supplementary Materials. Additional data related to this paper may be requested from the authors.
View Abstract

Stay Connected to Science Advances

Navigate This Article